When industry veterans Mark Thirsk and Mike Corbett hung out the Linx Consulting shingle a few years back, their firm was a welcome addition to the electronic materials information value chain.
I've known Mark for awhile, going back to his days at Shipley when it was still Shipley and wasn't part of Rohm and Haas. The tall Brit's balanced biz-tech knowledge of the lithography and CMP materials spaces, as well as his sly sense of humor have always been much appreciated.
Linx has just published a new report, Advanced Patterning: 2006-2011. They see the global photoresist, ancillaries, and related materials market pushing past the $1.5 billion mark this year and surpassing the $2 billion milestone in 2011. Overall growth during the next five years will be about 9%, but 193-nm resists, led by the dominant Japanese companies JSR and TOK, will grow at a 20% clip. The BARC [backside antiflective coating] and TARC [topside antireflective coating] 193-nm segments will see even stronger upticks, popping at 29% and 55%, respectively. But the healthiest jump of all will be experienced in the emerging multilayer-patterning segment, where the compound annual growth rates for new materials such as underlayers and spin-on hardmasks will bounce well in excess of 100-150%, according to Linx.
I emailed Mark some questions about the report and the advanced patterning materials market in general. Here are his replies.
What are the most active parts of the patterning materials market, and why?
193-nm resists are growing fast and are a significant market today. The older resist segments of i-line and 248 are maturing and will peak in the next 2 to 3 years. The most exciting segment, although growing from a small base, are materials for multilayer patterning. These include spin-on hardmasks and underlayers.
What are some of the most surprising findings in your study?
There is probably more uncertainty about future roadmaps for patterning now that at any time in the past. Despite the sums of money spent on EUV [extreme ultraviolet] and other NGL [next-generation lithography] technologies, they will struggle mightily to be ready for 32-nm manufacturing insertion. In fact, imprint lithography is making great progress toward overcoming its challenges.
There is no robust solution for 32 nm today. Immersion 193 needs to be augmented by aggressive DFM [design for manufacturability] and double-patterning schemes to be viable. Extensions of 193 immersion to higher refractive indices are facing some significant challenges too and will not be ready for 32-nm HVM [high-volume manufacturing] in 2010/2011.
Why have the Japanese companies (JSR and TOK) emerged as market leaders, especially in positive resists? Why have the second-tier companies (Fujifilm Electronic Materials [FFEM], Rohm and Haas Electronic Materials [RHEM], et al.) slipped? Can anyone, such as Dongjin, challenge the Japanese companies' dominance?
JSR and TOK have implemented well and shown that they have the capability to make competitive products at each new generation. Strong customer relationships have facilitated these successes, but each company must be given credit for having capable products available on time. Nearly all other suppliers have now “missed†a technology. FFEM and Sumitomo struggled to field winning 248-nm resists, ShinEtsu entered at 248, and RHEM and AZEM [AZ Electronic Materials] have not fared well in 193. Dongjin has a challenge to gain global acceptance, but is seen as very much being at a competitive standard and not lagging.
What might be some logical or at least possible consolidation moves in the market?
ShinEtsu and Sumitomo are both relatively small players focused on resists only. If another resist maker were willing to buy market share, they would be possible targets. Brewer is a strong player in the ancillary materials space and would be a very attractive addition to an electronic materials supplier not making ARCs [antireflective coatings].
How much will the cost of R&D and tech development in general increase in the market in the 2006-2011? What implications does that have for the sector and the customer base?
The main drivers of cost are the high cost of equipment (either to buy or to access) to develop products, as well as the need to customize nearly every product to customers. A resist manufacturer can easily exercise a scanner 24/7 and would prefer to own the latest tool months after its release to develop optimized products. With depreciation costs mounting to $5 million per year for such a tool, only those suppliers with strong share and growing revenues can sustain the pace profitably. The days of blockbuster products that meet multiple customer needs are long gone, and small batches and custom raw materials will keep product costs high.
Which customers are driving the sector the hardest (i.e., driving them crazy), technologically and economically?
Each segment has its own set of tough requirements. A device manufacturer will look for four to five critical resists for a new generation, each with a unique set of requirements. Covering three or so key customers leads to a possible development load of 12 to 15 consecutive resist development efforts.
Conversely, the number of key decision-makers (who define real volume resist/material purchases) is going down, making those resist evaluations very important. That doesn’t mean that gaining business at smaller customers is any easier---just a few weeks later.
What are some of the key emerging market segments, and why? Any new or up-and-coming players to watch?
Silicon hardmasks and soluble BARCs are interesting. Dow Corning has great silicon-containing polymer expertise and are supporting both RHEM and TOK product development. Linked to this, there are not too many companies supplying planarizing underlayers. These polymers are not uncomplicated and may be a big opportunity.
AZ and Brewer have interesting developable BARC products. If these products hit, they will be used in the implant levels, which are a third of all mask levels and rising in advanced devices. Despite challenges, these materials could be very big.
Along those same lines, what novel technologies have the best growth or opportunity potential, and why?
Multilayer products appear to be the area with the highest growth opportunity. Some of these materials will be needed in double-patterning schemes, and resists cannot provide the functionality that they used to, since imaging and high etch resistance in an ultrathin film are very difficult.
What do you see as the frontrunner technology for 32-nm HP from an advanced patterning materials perspective?
The ideal approach is a trilayer-type process with either a silicon-containing BARC, combined with a matched underlayer, or a spin-on hard mask matched with an organic BARC. Applied Materials and ASMI amorphous carbon films appear to be best suited to FEOL [front-end-of-line] levels, but are cheap, controllable layers if the chamber capacity is available.
Does immersion litho, combined with multiexposure and DFM bells and whistles, have a shot at being extendable to 22-nm HP?
That's the $64,000 question. There will probably need to be significant layout compromise to push optical beyond 32 nm. In prototyping I believe there will be little choice, but HVM may be a significantly different matter.
One constraint we have not touched on is LER/LWR [line-edge roughness/linewidth reduction]. It is probable that chemically amplified resists will hit the wall on LER at about 2-3 nm. Amorphous polymers, controlled PAG [photoacid generator] distribution, and increased dose to reduce shot noise all have a part to play, but there are very few demonstrations of LER below 3-nm 3 sigma, which has already eaten the 32-nm CD budget. A high-speed direct imaging resist (one photo = one photo event) is probably impractical, but if it is found, it may take 5 to 7 years to ramp to manufacture---too late for 32 nm and 22 nm.
|
Blogs 
